The One-Time Bill: America Is Paying Once to Stop Renting the World’s Cheap Labor

The tariff price shock is a step, not a slope — a single adjustment back to honest prices after thirty years of consumption subsidized by foreign labor conditions no American would accept at home. The payoff, unlike the bill, compounds.

The Hamiltonian Turn — Article II of a Vector1 Research Program

By Brian Curry · Vector1 Research

Article I argued that between 2018 and 2026 America reassembled Hamilton’s founding program. This paper takes on the loudest objection to the most contested instrument — “tariffs raised my prices” — and answers it with a decomposition the complaint never performs. The price increase is real; the federal data on it is unambiguous, and we present it in full. But it is a one-time level adjustment — a single step back to honest prices after three decades in which American consumption ran on a discount extracted from foreign workers laboring under conditions no American would tolerate for themselves. A step is paid once. The capability, alliance, and security gains purchased with it compound*. The ten-year arithmetic between a flat bill and a compounding payoff is not close — and the discomfort of the adjustment is not an argument against it, any more than the first month off a borrowed credit card is an argument for keeping the card.*

1. What the Complaint Is Actually About

Start with the chart that explains thirty years of American consumer life. From the mid-1990s to 2020, the price of durable goods in America fell — televisions, furniture, appliances, electronics, clothing, decade after decade, cheaper in nominal terms while every domestically produced thing (housing, healthcare, education, services) marched upward. Economists called it the great goods deflation. It deserves a more honest name: the imported subsidy.

Figure 1. Durable-goods prices versus services prices since 1995. The divergence is not a law of nature; it is the offshoring arbitrage made visible. Goods got cheaper because their production was moved to places where labor is paid a fraction of American wages, organized under conditions ranging from harsh to — in documented cases — forced. The 2025–26 step is the line beginning its return toward true cost.

That falling red-blue gap financed a particular way of life: the highest consumption share of GDP in the developed world, a personal savings rate that fell from a 10% historical norm to the low single digits, and a national habit of treating a $40 toaster and a $1,200 flagship phone as natural prices rather than as what they are — prices net of a subsidy paid by someone else, somewhere else. When tariffs step prices back toward cost, the resulting complaint is genuine and politically potent. But decompose it and the complaint is not, at bottom, about tariffs. It is about the end of the discount — and an honest accounting must say what the discount was made of.

2. What the Discount Was Made Of

Write the consumption identity of the offshoring era explicitly:

American consumption ran above American income on two inputs: a collapsing savings rate at home, and an imported subsidy from abroad. And the imported subsidy itself decomposes into three parts, none of which belongs in an honest price:

The first term is repressed labor cost — and this is not rhetoric; it is enforcement record. The Uyghur Forced Labor Prevention Act exists because forced labor was documented inside these supply chains, and U.S. Customs has detained billions of dollars of goods under it. Beyond the outright forced-labor cases sits the broader structure: workers without independent unions, without meaningful labor law, in dormitory-factory systems whose conditions periodically produce scandals that briefly trouble Western consciences before the next product launch. The second term is state subsidy — capital, land, energy, and currency arrangements engineered to take strategic industries. The third is unpriced coercion risk — the exposure that stopped being theoretical when export controls on rare earths were actually imposed in 2025.

Here is the uncomfortable sentence the public debate avoids, so this paper will say it plainly: a consumer who demands the pre-2025 price level is demanding the continuation of all three terms. The discount was never free. It was paid — by workers with no recourse, by the hollowed industrial regions that absorbed the arbitrage at home, and by a strategic exposure the country is only now unwinding. Cheap was a place, and a set of conditions, that Americans would not accept for one hour of their own labor. The moral arithmetic does not change because the savings were convenient.

3. The Bill, Stated Honestly

None of this argues for pretending the adjustment is painless, and this series does not launder data. The Federal Reserve’s real-time work finds pass-through of the 2025 tariffs to consumer prices is full — dollar-for-dollar, complete roughly seven months after implementation — raising core goods prices about 3.1% and core PCE about 0.8%, roughly $1,500 per household per year at current scope:

Figure 2. Who pays, and when. The bill is real, it is regressive in the short run — goods are a larger share of lower-income budgets — and any honest defense of the policy owns it rather than disputes it. This paper’s argument is not that the bill is fake. It is that the bill is singular*.*

The distributional point deserves its own sentence: the step lands hardest on households least able to absorb it, which is a real cost and a real political constraint — and the strongest argument for routing the compounding gains (the jobs at $30 an hour, the dividend mechanisms of our companion series) visibly back to the same households. But the size and incidence of the bill are not the question that decides the policy. The question is its shape.

4. The Theorem: A Step Is Not a Slope

Here is the center of this paper, and it is one line of mathematics. A tariff raises the price level, once. A learning curve lowers production cost at a compounding rate, forever:

These are different mathematical species, and confusing them is the central error of the entire price debate. Inflation is a rate — the change in the level. A one-time step enters the inflation statistics for exactly one year and then falls out of them arithmetically, because a step has no slope:

This is not a forecast; it is what the word “rate” means — and it is why the Fed’s own framing treats the tariff effect as a one-time level shift to look through rather than an inflation process to fight, provided expectations hold. Meanwhile the other side of the ledger is governed by Wright’s law: every doubling of cumulative domestic output cuts unit cost 20–30% in chips and batteries, the gains compounding annually, never reversing, never falling out of any base year:

Figure 3. The whole argument in one picture. The price level steps up ~0.8% once and goes flat. Capability compounds and never stops. Any evaluation horizon longer than about three years is dominated by the slope; the public debate has been conducted entirely inside the step.

The complaint about prices, in other words, is an argument about the first eighteen months of a thirty-year compounding process — the economic equivalent of judging a mortgage by the closing costs.

5. What the Step Already Bought

The compounding side is not hypothetical; four years of it are now on the books, and Article I and the record since supply the entries. Capability: $33 billion in CHIPS awards has unlocked roughly $645 billion in announced private investment across thirty states; 97 fabs launched in three years; TSMC’s Arizona yields competitive with Taiwan; ~71,000 chip-ecosystem jobs at factory wages around $30 an hour; manufacturing PMI at a four-year high in May 2026 — with the honest footnote that the PMI surge is driven substantially by the AI buildout and Iran-war front-loading, and that factory employment has still contracted for 32 straight months. The learning dividend these purchases compound at:

De-risking: China’s share of U.S. imports cut from ~22% to ~7% — direct exposure to a supplier that demonstrated its willingness to weaponize trade, reduced by two-thirds, the fastest peacetime reorientation of a major economy’s supply base on record (with the connector-country caveat honestly attached):

Leverage: the NATO Hague commitment — thirty-two allies to 5% of GDP, roughly $375 billion a year of additional allied defense spending at maturity, which the alliance’s own Secretary-General said would not have happened without American pressure — plus the reciprocity corrections and framework pledges:

Figure 5. What the bargaining position bought, against its carrying cost. Attribution is shared and pledges are soft — the haircuts are in the text — but the order of magnitude survives them.

6. The Ten-Year Test

Now put both sides on one axis and run the test this series was built on — the Mill–Bastable present-value condition, evaluated over the horizon that actually matters:

Figure 6. Cumulative cost against cumulative benefit, 2026–2036. The bill flattens — by definition, because a level shock is paid once. The benefits compound — by Wright’s law, by alliance budget phase-ins, by every crisis the de-risked supply chain doesn’t become. The crossover lands around the turn of the decade and is robust to ±50% haircuts on every benefit line. This is what “short-term pain, long-term gain” looks like when you actually do the math instead of saying the phrase.

This is the answer to the evaluation-horizon question. Graded on eighteen months of price data, the tariff instrument looks like a tax, because in eighteen months that is mostly what any investment looks like — outlay first, returns later. Graded on the decade, the step purchases a compounding stream that dwarfs it. The error is not in the Fed’s data — the data is the cost side of the integral, correctly measured. The error is in mistaking the cost side for the verdict.

7. The Instruments, Still Ranked

None of this suspends Article I’s discipline, and the scorecard survives the reframe intact — because the reframe is about horizon, not about targeting. Subsidies aimed at steep learning curves still dominate (A−); the equity-plus-offtake model still prices security best (A−); and the broad tariff still carries its split grade — weak as protection, strong as leverage and de-risking — with the same policy implication: migrate it toward statutory, China-and-conduct-calibrated form, where the true-cost wedge is real, and off allies, where it is not. The one-time bill buys the most when the step lands precisely on the prices that were dishonest, and least when it taxes Canadian lumber at the rate meant for forced-labor supply chains:

Figure 7. The instrument ranking, unchanged by the horizon argument: targeting still decides the grade. The ten-year math vindicates the turn; it does not vindicate every tool equally.

8. Objections, Taken Seriously

“You’re rationalizing a tax on the poor as a moral awakening.” The regressivity is conceded in §3, in print, with the policy answer attached: route the compounding gains visibly to the households paying the step — the manufacturing wage premium, and the dividend architectures of our companion series. What we decline to concede is that the pre-2025 arrangement was the pro-poor position: a discount financed by forced and repressed labor abroad and by the destruction of the domestic working class’s industrial base was not a progressive policy; it was a regressive policy with good prices.

“Prices won’t step once — tariff scope keeps changing, so the ‘one-time’ claim is convenient.” Fair as a critique of policy volatility, not of the arithmetic. Each scope change is its own step; a sequence of steps is still not a compounding rate, and the base-year algebra applies to each. The volatility argument is the strongest case for the statutory calibration §7 recommends — make the step once, precisely, and stop.

“The ISM rebound proves protection is working / the employment data proves it isn’t.” Both readings overreach, and §5 prints both facts side by side: a four-year-high PMI driven substantially by AI capex and war front-loading, and a 32nd straight month of employment contraction. The compounding case does not rest on either — it rests on realized investment, yields, and learning rates, which are measured and large.

“Ten-year projections can justify anything.” Which is why Figure 6 publishes its components and survives ±50% haircuts, why the bill side uses the Fed’s own numbers unshaded, and why the crossover claim is conditional on the phase-ins actually occurring — the operating-cost convergence named in Article I remains the live empirical test. A projection with published assumptions is an argument; we have made one. The alternative on offer — judging a capability program by its closing costs — is not a conservative method; it is a category error with a green eyeshade.

“You’re calling American consumers complicit in forced labor.” We are saying the price contained it, the enforcement record proves it, and that demand for the old price level is demand for what produced it — knowingly or not. The point of the adjustment is precisely to let Americans buy goods whose prices contain no one’s coercion. That is not an accusation. It is an invoice, finally itemized.

9. What Remains

The diagnosis (Article I) said America reverted to its founding program. This paper said the program’s most resented cost is a one-time step back to honest prices — and its gains compound. What remains is the institution that decides whether the compounding is captured or squandered: the de facto sovereign portfolio, the full construction of the Hamilton Index, and the 2050 projection with its three failure modes. Article III — The American Fund and the Hamiltonian Century — closes the series.

References and Sources

1. BLS Consumer Price Index, durable goods vs. services series (1995–2026) — the divergence of Figure 1.
2. Minton, R., Ray, M., & Somale, M. (April 2026). Federal Reserve FEDS Notes. Full dollar-for-dollar pass-through ~7 months; +3.1% core goods; +0.8% core PCE.
3. Tax Foundation (2026): ~$1,500 average household cost. NY Fed (Feb 2026): ~90% borne domestically.
4. Uyghur Forced Labor Prevention Act (2021); CBP detention statistics — the documented forced-labor basis of the true-cost wedge.
5. SIA, Chip Supply Chain Investments (2026): ~$645B announced private investment across 30 states; ~71,000 ecosystem jobs; $33B+ in awarded grants. SEMI: 97 fabs (2023–25).
6. ISM Manufacturing PMI Report (June 1, 2026): PMI 54.0, highest since May 2022; New Orders 56.8; Employment 32nd month of contraction. Reuters (June 1, 2026): expansion “anchored mostly by an artificial intelligence spending spree”; Iran-war front-loading; manufacturing employment −77,000 since Jan 2025.
7. NATO, The Hague Summit Declaration (2025): 5% of GDP by 2035. Defense News: Rutte credits U.S. pressure.
8. Census/CEPR/Euromonitor/Deutsche Bank (2024–26): China import share 22% → 13.1% → ~7%; imports from China $536B → $308B; connector-country intermediation caveat.
9. Wright, T. P. (1936); Juhász, Lane & Rodrik (2023) — learning-curve rates and the industrial-policy evidence base.
10. Jones & Marinescu and companion literature via Vector1, Returns to Data V — the dividend-routing mechanisms referenced in §3 and §8.
11. Vector1 Research (2026). The Hamiltonian Turn, Article I: the framework, Equation 4’s targeting discipline, the Hamilton Index.

Methodological note: Figures 1 and 3 are stylized renderings of the cited series and mechanisms; Figure 6 is a scenario whose components (learning dividend at Wright-law rates on realized investment; NATO uplift at (3.5%−2%)×allied GDP with phase-in; dependency value as avoided-coercion expected loss) are stated so each can be attacked, and whose crossover survives ±50% haircuts on every benefit line. The bill side uses the Fed’s pass-through estimates without adjustment. The regressivity of the step and the 32-month employment contraction are reported unshaded in §3 and §5. The “one-time” characterization is conditional on scope stabilization, as §8 states. Dollar figures nominal. Code: github.com/vector1-ai (forthcoming).

Brian Curry is a Kansas City–based data scientist, AI researcher, and founder of Vector1 Research. The Hamiltonian Turn program develops the economic theory and measurement of America’s return to developmental statecraft.

Contact: brian at vector1.ai · License: CC BY 4.0

#HamiltonianEconomics #Tariffs #TruePrices #Reshoring #ForcedLabor #SupplyChains #Manufacturing #LearningCurves #TradePolicy #Vector1Research